Betas calculated purely based on historical data are unadjusted betas. However, this beta estimate based on historical estimates is not a good indicator of the future. This is also called the beta instability problem.

Statistically, over time betas may exhibit mean reverting properties as extended periods significantly above 1 (one) may eventually decline and betas below one may revert toward 1.

Therefore analysts may apply models to create an adjustment calculation for the historical beta and use this adjusted beta to calculate the expected return for the security.

The generalized formula for adjusted beta can be presented as follows: